La start-up Anthropic has given access to Spain a Mythosa mysterious model of artificial intelligence that has put the world of cybersecurity on alert, as confirmed by sources from the Government to EL PERIÓDICO.
Last April, the emerging company announced that it had created a IA capable of exposing vulnerabilities of computer systems with unprecedented speed and precision, as well as execute cyber attacks sophisticated attacks against these critical infrastructures. Although several experts have disputed Anthropic’s account, they suggest that Mythos is a very potent offensive weapon.
Those capabilities led Anthropic to rule out a public launch and restrict its access to a small group of companies. Still, they announced it with great fanfare.
Access to Europe
Concern about the hypothetical impact of that system led to a flood of access requests to Anthropic, including from the European Union. The Spanish Vice President and Minister of Economy, Carlos Bodyasked in early May to have “early access” to that tool to “not be a second-class region.”
Pressure from Brussels has borne fruit. Anthropic announced on Monday that it had included the European Union Agency for Cybersecurity (ENISA) in the so-called ‘Project Glasswing’, the select group of 150 institutions and companies from more than 15 countries that have been allowed access to Mythos Preview for “defensive purposes.” That is, in order to anticipate the risks and opportunities it poses.
Dario Amodei, co-founder and CEO of Anthropic, the AI company behind the Claude models. / Ruhani Kaur / Bloomberg
In this way, the invitation to the EU is extended to Spain. Participating in the testing of this model, sources from the Executive assure, “is an example of the influence of the Government of Spain in the debate on frontier AI governance” that “consolidates its international leadership in responsible AI development.”
Anthropic adelanta a OpenAI
The firm led by Dario Amodeiwhose valuation is close to a trillion dollars, appears to be outperforming in several metrics OpenAIthe creator of ChatGPT. Thanks to its model of Generative AInicknamed ClaudeAnthropic has overtaken its main rival in enterprise customers and expects its first operating profit of $559 million for the second quarter. OpenAI, on the other hand, projects losses annually of 14,000 million dollars. To make matters worse, Anthropic has also overtaken the company from which it was spun off in its procedures to go public, an operation that could be the largest ever recorded.
Subscribe to continue reading
